AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Alpha-divergence Variational Inference Meets Importance Weighted Auto-Encoders: Methodology and Asymptotics

Kamélia Daudel, Joe Benton, Yuyang Shi, Arnaud Doucet

NeurIPS 2023 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Amortized and online variational inference · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Increase dimension d with fixed N (log-normal weights, B_d = sqrt(d))VR-IWAE bound · Gaussian example, d in {10,100,1000}variational gapHurt
Train phi so that B_d/sqrt(d) decreases from 1 to 0VR-IWAE bound · Gaussian example, alpha=0.2, d=1000variational gap approximation by fixed-d 1/N asymptoticsImproved
Increase alpha in [0,1) for the boundVR-IWAE bound · Gaussian exampleweight-collapse limitation in high dimensionNo clear effect
Use alpha > 0 instead of alpha = 0 (IWAE)reparameterized inference network gradient · Linear Gaussian, d in {20,100}, sigma_pert=0.01SNR of inference network gradientImproved
Increase alphaVR-IWAE gradient estimators · Linear Gaussian, low-medium dSNR of generative and inference networksImproved
Move parameters away from optimum (sigma_pert=0.5) or increase d to 1000SNR gain from alpha > 0 · Linear GaussianSNR improvementHurt
Use doubly-reparameterized gradient estimator for phiVR-IWAE inference network gradients · Linear GaussianSNR decayImproved
Increase alphalog-likelihood estimation · Linear GaussianMSE of log-likelihood estimateNo clear effect
Increase alphatheta gradient estimates · Linear GaussianMSE of theta gradientImproved
Use optimal parameters (sigma_pert=0) with increasing dVR-IWAE bound · Linear GaussianVR-IWAE boundHurt
Let N grow faster than d^(1/3)asymptotic regime of the bound · Linear Gaussianvalidity of the high-dimensional asymptoticHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy